Add operation label to user metric and move from manager to api

This adds an "operation" label to the user metrics and moves incrementing from the manager to api, so that the user metrics show when resources are allocated, as well as destroyed. Previously, user metrics were only updated upon destroying a resource.

I think its better suited to increment the metric as part of the api instead of the pool_manger, because it's expected to do so when a user successfully checks out or deletes a VM, but can be problematic when doing so in the provider since it can clone VMs before actually being checked out by a user.

lib/vmpooler/api/v1.rb

@@ -210,6 +210,7 @@ def atomically_allocate_vms(payload)
             account_for_starting_vm(vmpool, vmname)
             vms << [vmpool, vmname, vmtemplate]
             metrics.increment("checkout.success.#{vmpool}")
+            update_user_metrics('allocate', vmname) if Vmpooler::API.settings.config[:config]['usage_stats']
           else
             failed = true
             metrics.increment("checkout.empty.#{requested}")
@@ -235,6 +236,46 @@ def atomically_allocate_vms(payload)
       result
     end
 
+    def update_user_metrics(operation, vmname)
+      backend.multi
+      backend.hget("vmpooler__vm__#{vmname}", 'tag:jenkins_build_url')
+      backend.hget("vmpooler__vm__#{vmname}", 'token:user')
+      backend.hget("vmpooler__vm__#{vmname}", 'template')
+      jenkins_build_url, user, poolname = backend.exec
+
+      if user
+        user = user.gsub('.', '_')
+      else
+        user = 'unauthenticated'
+      end
+      metrics.increment("user.#{user}.#{operation}.#{poolname}")
+
+      if jenkins_build_url
+        if jenkins_build_url.include? 'litmus'
+          # Very simple filter for Litmus jobs - just count them coming through for the moment.
+          metrics.increment("usage_litmus.#{user}.#{operation}.#{poolname}")
+          return
+        end
+
+        url_parts = jenkins_build_url.split('/')[2..-1]
+        jenkins_instance = url_parts[0].gsub('.', '_')
+        value_stream_parts = url_parts[2].split('_')
+        value_stream_parts = value_stream_parts.map { |s| s.gsub('.', '_') }
+        value_stream = value_stream_parts.shift
+        branch = value_stream_parts.pop
+        project = value_stream_parts.shift
+        job_name = value_stream_parts.join('_')
+        build_metadata_parts = url_parts[3]
+        component_to_test = component_to_test('RMM_COMPONENT_TO_TEST_NAME', build_metadata_parts)
+
+        metrics.increment("usage_jenkins_instance.#{jenkins_instance}.#{value_stream}.#{operation}.#{poolname}")
+        metrics.increment("usage_branch_project.#{branch}.#{project}.#{operation}.#{poolname}")
+        metrics.increment("usage_job_component.#{job_name}.#{component_to_test}.#{operation}.#{poolname}")
+      end
+    rescue StandardError => e
+      puts 'd', "[!] [#{poolname}] failed while evaluating usage labels on '#{vmname}' with an error: #{e}"
+    end
+
     def update_pool_size(payload)
       result = { 'ok' => false }
 
@@ -1169,6 +1210,7 @@ def delete_ondemand_request(request_id)
           status 200
           result['ok'] = true
           metrics.increment('delete.success')
+          update_user_metrics('destroy', params[:hostname]) if Vmpooler::API.settings.config[:config]['usage_stats']
         else
           metrics.increment('delete.failed')
         end

lib/vmpooler/metrics/promstats.rb

@@ -165,33 +165,33 @@ def vmpooler_metrics_table
           },
           user: {
             mtype: M_COUNTER,
-            torun: %i[manager],
-            docstring: 'Number of pool instances this user created created',
-            param_labels: %i[user poolname]
+            torun: %i[api],
+            docstring: 'Number of pool instances and the operation performed by a user',
+            param_labels: %i[user operation poolname]
           },
           usage_litmus: {
             mtype: M_COUNTER,
-            torun: %i[manager],
-            docstring: 'Pools by Litmus job usage',
-            param_labels: %i[user poolname]
+            torun: %i[api],
+            docstring: 'Number of pool instances and the operation performed by Litmus jobs',
+            param_labels: %i[user operation poolname]
           },
           usage_jenkins_instance: {
             mtype: M_COUNTER,
-            torun: %i[manager],
-            docstring: 'Pools by Jenkins instance usage',
-            param_labels: %i[jenkins_instance value_stream poolname]
+            torun: %i[api],
+            docstring: 'Number of pool instances and the operation performed by Jenkins instances',
+            param_labels: %i[jenkins_instance value_stream operation poolname]
           },
           usage_branch_project: {
             mtype: M_COUNTER,
-            torun: %i[manager],
-            docstring: 'Pools by branch/project usage',
-            param_labels: %i[branch project poolname]
+            torun: %i[api],
+            docstring: 'Number of pool instances and the operation performed by Jenkins branch/project',
+            param_labels: %i[branch project operation poolname]
           },
           usage_job_component: {
             mtype: M_COUNTER,
-            torun: %i[manager],
-            docstring: 'Pools by job/component usage',
-            param_labels: %i[job_name component_to_test poolname]
+            torun: %i[api],
+            docstring: 'Number of pool instances and the operation performed by Jenkins job/component',
+            param_labels: %i[job_name component_to_test operation poolname]
           },
           checkout: {
             mtype: M_COUNTER,

lib/vmpooler/pool_manager.rb

@@ -473,54 +473,11 @@ def _destroy_vm(vm, pool, provider)
           finish = format('%<time>.2f', time: Time.now - start)
           $logger.log('s', "[-] [#{pool}] '#{vm}' destroyed in #{finish} seconds")
           $metrics.timing("destroy.#{pool}", finish)
-          get_vm_usage_labels(vm, redis)
         end
       end
       dereference_mutex(vm)
     end
 
-    def get_vm_usage_labels(vm, redis)
-      return unless $config[:config]['usage_stats']
-
-      redis.multi
-      redis.hget("vmpooler__vm__#{vm}", 'checkout')
-      redis.hget("vmpooler__vm__#{vm}", 'tag:jenkins_build_url')
-      redis.hget("vmpooler__vm__#{vm}", 'token:user')
-      redis.hget("vmpooler__vm__#{vm}", 'template')
-      checkout, jenkins_build_url, user, poolname = redis.exec
-      return if checkout.nil?
-
-      user ||= 'unauthenticated'
-      user = user.gsub('.', '_')
-      $metrics.increment("user.#{user}.#{poolname}")
-
-      return unless jenkins_build_url
-
-      if jenkins_build_url.include? 'litmus'
-        # Very simple filter for Litmus jobs - just count them coming through for the moment.
-        $metrics.increment("usage_litmus.#{user}.#{poolname}")
-        return
-      end
-
-      url_parts = jenkins_build_url.split('/')[2..-1]
-      jenkins_instance = url_parts[0].gsub('.', '_')
-      value_stream_parts = url_parts[2].split('_')
-      value_stream_parts = value_stream_parts.map { |s| s.gsub('.', '_') }
-      value_stream = value_stream_parts.shift
-      branch = value_stream_parts.pop
-      project = value_stream_parts.shift
-      job_name = value_stream_parts.join('_')
-      build_metadata_parts = url_parts[3]
-      component_to_test = component_to_test('RMM_COMPONENT_TO_TEST_NAME', build_metadata_parts)
-
-      $metrics.increment("usage_jenkins_instance.#{jenkins_instance}.#{value_stream}.#{poolname}")
-      $metrics.increment("usage_branch_project.#{branch}.#{project}.#{poolname}")
-      $metrics.increment("usage_job_component.#{job_name}.#{component_to_test}.#{poolname}")
-    rescue StandardError => e
-      $logger.log('d', "[!] [#{poolname}] failed while evaluating usage labels on '#{vm}' with an error: #{e}")
-      raise
-    end
-
     def component_to_test(match, labels_string)
       return if labels_string.nil?